Real-time windrow detection from onboard tractor sensors for automated following
This work provides an open-source benchmark and dataset for GPS-free windrow detection, addressing the lack of transparency in commercial systems for autonomous forage-harvesting research.
The authors present a multi-modal dataset (stereo vision + LiDAR) for windrow detection in forage harvesting and implement a real-time (>20 Hz) centroid-based following method on an NVIDIA Jetson AGX Orin. Stereo and LiDAR depth measurements show strong agreement (0.965 ± 0.021) in the 4-10 m range, suggesting low-cost stereo can approach LiDAR performance.
Proprietary design in commercial windrow-detection systems restricts transparency and limits progress in open autonomous forage-harvesting research. We present a multi-modal dataset combining stereo vision and LiDAR from tractor-mounted sensors during real baling operations. The dataset includes synchronized sensor data with GNSS trajectories, partly released as ROS2 Humble bags on Zenodo, with additional data available on request. Using this dataset, we implement a real-time (>20 Hz) centroid-based windrow-following method on an NVIDIA Jetson AGX Orin. Across the critical 4-10 m guidance range, stereo and LiDAR depth measurements show strong agreement (0.965 +/- 0.021), indicating that low-cost stereo sensors can approach LiDAR performance. Our open-source ROS 2 pipeline provides a reproducible benchmark for GPS-free windrow detection and supports development of practical autonomous forage-harvesting systems. Dataset: https://zenodo.org/records/17486318